| name | research-project-management |
| description | Work plan tracking, dependency management, milestone planning, reviewer response strategy. Use when organizing research timelines, tracking experiment progress, planning thesis or paper milestones, or preparing reviewer rebuttals. |
Research Project Management
Follow these practices when planning, tracking, and managing academic or applied research projects.
1. Work Plan Structure
Decompose the project into phases with clear deliverables:
Phase 1: Foundation (Weeks 1–3)
├── Literature review complete
├── Baseline implementation verified
└── Evaluation pipeline tested end-to-end
Phase 2: Core Contribution (Weeks 4–8)
├── Proposed method implemented
├── Initial results on primary benchmark
└── Ablation study designed
Phase 3: Evaluation & Writing (Weeks 9–12)
├── Full experimental campaign (multi-seed, all benchmarks)
├── Paper draft complete
└── Figures finalized (camera-ready quality)
- Each phase should have entry criteria (what must be true before starting) and exit criteria (what must be delivered to move on).
- Track progress with a simple checklist or kanban board, not heavyweight project management tools.
2. Dependency Management
Research tasks often have hidden dependencies. Map them explicitly:
- Data dependencies: Which experiments require which datasets to be prepared?
- Compute dependencies: Which experiments need GPU access, cluster time, or long-running slots?
- Knowledge dependencies: Which experiments require results from earlier experiments to inform design decisions?
- External dependencies: Waiting on access to a system, a collaborator's code, or a dataset license?
Identify the critical path — the longest chain of sequential dependencies — and prioritize unblocking it.
3. Milestone Planning
Align milestones with external deadlines:
| Milestone | Typical Deadline |
|---|
| Paper submission | Conference deadline (hard) |
| Thesis chapter draft | Supervisor review (soft) |
| Experiment completion | 2–3 weeks before writing begins |
| Camera-ready figures | 1 week before submission |
| Code cleanup & release | Before or at publication |
- Buffer rule: Plan to finish experiments at least 2 weeks before the writing deadline. Writing always takes longer than expected.
- Parallel tracks: While experiments run (often overnight or over days), use that time for writing, literature review, or code cleanup.
4. Experiment Tracking Integration
- Maintain a running experiment log (markdown, spreadsheet, or notebook) that records:
- What was tried (hypothesis, configuration)
- What happened (result, observation)
- What was learned (insight, next step)
- Tag experiments as:
exploratory, confirmatory, or failed (with reason).
- Never delete failed experiments — they inform future decisions.
5. Reviewer Response Strategy
When responding to peer reviews:
- Acknowledge every point: Even if you disagree, show you understood the reviewer's concern.
- Be concrete: "We added experiment X (Table 3) which shows Y" is better than "We addressed this concern."
- Diff the paper: Provide a marked-up version showing exactly what changed.
- Prioritize: Address major concerns first. Minor formatting or typo fixes can be grouped.
- Stay respectful: Reviewers volunteer their time. Thank them, even for harsh reviews.
6. Collaboration Practices
- Use version control for everything: code, paper source (LaTeX), and experiment configurations.
- Establish a shared naming convention for experiments, branches, and result directories.
- Schedule regular syncs (weekly for active projects) with a fixed agenda: progress, blockers, next steps.
- Keep a shared "decisions log" for design choices made during meetings, so they don't get lost.
